Building the Trust Layer for the Agentic Era | Sonar Summit 2026

NOTE: The roadmap has changed since we recorded this presentation. How do you build trust in AI-generated code? In this Sonar Summit 2026 keynote, Ori explores the growing trust gap in the AI-native SDLC as development teams adopt AI assistants and autonomous agents to generate code at unprecedented speed. While AI accelerates software development, it also increases the risk of introducing bugs, security vulnerabilities, and technical debt faster than teams can review them. As organizations move from simple AI copilots to agentic AI systems, traditional code review processes struggle to keep up with the scale and complexity of AI-generated changes. In this talk, you’ll learn: Why AI-assisted development creates a verification challenge How agentic AI changes the software development lifecycle The growing gap between code generation and code verification Why organizations need a trust layer for AI-driven development Sonar’s vision for establishing SonarQube as the verification layer for code quality and security Discover how teams can maintain high standards for software quality, application security, and secure coding practices while enabling developers and AI agents to collaborate at scale. By integrating automated static analysis, vulnerability detection, and software verification into the development workflow, organizations can maintain control over rapidly growing codebases while embracing the productivity benefits of AI-powered development. Timestamps: 00:00 — Introduction 00:17 — Adapting Faster in the Era of AI-Powered Development 01:01 — Copilot Coding vs Agent-Driven Development Models 01:23 — Why AI Coding Agents Need Planning and Context 02:07 — The AI Code Generation and Verification Loop 02:58 — The Guide Stage: Providing Context to AI Coding Agents 03:37 — Context Augmentation for AI Software Development Agents 04:05 — What “Context” Means in AI-Assisted Coding 04:42 — Delivering Code Context to AI Developer Tools 05:09 — Static and Dynamic Context for AI Code Generation 05:57 — Benchmark Results: Fewer Bugs and Vulnerabilities with Context 06:43 — MCP Server: Integrating AI Coding Agents with Developer Tools 07:30 — Architecture Drift in AI-Generated Code 08:14 — Introducing SonarQube Architecture for Governance 09:28 — Architecture Discovery and Codebase Dashboards 10:01 — Defining Intended Architecture and Enforcing Rules 10:28 — Detecting Architecture Deviations and Code Quality Issues 10:53 — Guide Stage Recap and Product Roadmap Updates 11:26 — Verify Stage: Checking AI-Generated Code with Analysis 12:12 — Agentic Code Analysis for Faster Verification 12:49 — Running Sonar Verification in the Developer Workflow 13:30 — Sending Code Quality Feedback to AI Coding Agents 14:43 — Creating Clean Pull Requests with Automated Verification 15:11 — Solve Stage: AI Remediation Agents Fixing Code Issues 15:54 — Remediation Agent Benchmark Results 16:50 — Enabling and Assigning a Code Remediation Agent 17:48 — Monitoring AI Agent Fixes and Pull Request Changes 18:27 — Automatic Code Remediation After Failed Quality Gates 19:08 — Guide–Verify–Solve Workflow for AI Development 19:28 — Final Thoughts: Responsible AI-Assisted Software Development #SonarSummit #AIinSoftwareDevelopment #DevSecOps #ApplicationSecurity #SoftwareQuality